To create a new partitioned variable, click + Create partitioned variable. New partitioned variables can be persisted in the output table, and also used in additional queries in the partition section.
Within the function definition, you'll choose one or more variables to Partition on, which defines a set of partitions, each of which contains the rows containing a unique combination of values across those variables.
For example, we can partition on the
provider_state variables to create a set of partitions, each containing the rows with a given unique combination of
provider_state. Our partitioned variable
average_reimbursement_per_city computes a mean of
average_remibursement across all rows in the partition and assigns that value to every row in the partition. So all records from
AZ are assigned a computed
average_reimbursement_per_city value, all records from
AZ are assigned a different computed
average_reminbursement_per_city value, and so on.
You can read more about partitioned variables here.